PUBLIC INVESTMENT IN TRANSPORTATION INFRASTRUCTURES
AND ECONOMIC PERFORMANCE IN PORTUGAL
Alfredo M. Pereira
Department of Economics
The College of William and Mary
Williamsburg, VA 23187
Email: [email protected]
Jorge M. Andraz
Faculdade de Economia
Universidade do Algarve, Campus de Gambelas
Faro, Portugal
Email: [email protected]
Abstract - This paper uses a VAR approach to investigate the effects of aggregate and disaggregate
measures of public investment in transportation infrastructures on private investment, employment, and output
in Portugal. Estimation results suggest that public investment in transportation infrastructures crowds in
private investment and employment and, therefore, has a strong positive effect on output. Indeed, we estimate
that one euro invested in public investment increases output in the long-term by 9.5 euros. This figure suggests
that public investment pays for itself 3.3 times in the form of tax revenues over the life span of the public
capital asset. Furthermore, this figure corresponds to a rate of return of 15.9%, which is clearly higher than
the rate of return expected on private investment activities. A close look at the effects of different types of
public investment is very informative, since it shows which types of public investments are the most
productive. In terms of marginal productivities, the highest effects on private investment come from public
investment in ports, airports and national roads. In terms of job creation, the highest effects come from public
investment in ports, municipal roads, and national roads. Finally, in terms of the effects on output the largest
effects come from investment in ports followed by national roads, municipal roads, airports, and railroads.
The results in this paper are very important from a public policy perspective. This is because they suggest that
public investment in transportation infrastructures has been a powerful instrument to promote long-term
growth and that the strategy followed by the Portuguese authorities of investing in public infrastructures is
justified both from a long-term development perspective as well as from a public budgetary perspective.
JEL Classification: C32, E62, H54, O52.
Keywords: infrastructures, economic performance, Portugal.
(*) This paper is part of a research project on the impact of infrastructure investment in the Portuguese
economy sponsored by the Fundação Luso-Americana/Portuguese-American Foundation. This paper was
prepared in the context of the conference on “Desenvolvimento Económico Português no Espaço Económico
Europeu: Determinantes e Políticas” organized by the Banco de Portugal/Bank of Portugal. We would like to
thank an anonymous referee for this conference for very useful comments and suggestions.
1
I. Introduction
One of the nagging aspects in the Portuguese economic performance is the relative backwardness of
the Portuguese economy vis-à-vis the European Union partners. From the 1970s until around the late 1980s
the Portuguese GDP per capita in purchasing power parity was just approximately around 55% of the EU
average. The magnitude and persistence of the relative backwardness of the Portuguese economy has been
explained by the lack of domestic long-term growth fundamentals. Historically serious distortions in the
financial markets lead to lagging private investment while a narrow domestic tax base hindered the
development of a modern infrastructure. These difficulties justified the EU structural funds programs after
1989. The cornerstone of these structural transfer programs was the development of a modern transportation
infrastructure network. Therefore, over the last decade, the strategy of development in Portugal has been
largely based on transportation infrastructure development.
Interestingly enough, despite the central role of public infrastructure development and the intuitive
knowledge of the relative scarcity of public infrastructures in Portugal, no information was available on the
actual impact of this development strategy. While the impact of infrastructure development on private
investment, employment and output has been assumed to be positive and important, there has been complete
ignorance as to what the actual effects might be. In particular, no estimates exist of the rates of return on
different types of infrastructure investment. Therefore no information exists on relationship between the rates
of return on public investments and the rate of return on private investment projects. This information,
however, is crucial to determine the appropriateness of the development strategy followed in Portugal.
The most important reason for the absence of estimates of the rates of return to public infrastructure
investment in Portugal, as indeed for most other countries of comparable or lower levels of development, has
been the absence of the most basic data on public investment itself. This is because of the highly
decentralized nature of the institutions in charge of the different types of public investment as well as the
constant shifting in jurisdictions in public investment activities. In the case of Portugal, however, the problem
of the absence of a data set has now been solved. The authors concluded recently the construction of a
detailed database of public investment in transportation infrastructures, under the auspices of the Portuguese
2
Ministry of Planning [see Pereira and Andraz (2001)].
In this paper we estimate empirically the impact of infrastructure investment in the area of
transportation on aggregate economic performance in Portugal. We focus on aggregate public investment as
well as on different types of transportation infrastructure - roads and highways, ports, airports, and railways to evaluate the effects of such public investments on private investment, employment and, ultimately, on
output. We seek to estimate the marginal products and the rates of return of public investment in different
types of transportation infrastructures.
Although this paper focuses on the Portuguese case and deals with issues that are of great importance
for policy making in Portugal, its interest is not merely parochial. Indeed, the issue of the effects of public
investment on private sector performance has been at the center of the policy debate in many countries, in
many regions of the world. In particular, in the European Union, the development strategy of the less
development countries, like Greece, Ireland, and Portugal, has been based largely on public investment
projects. For these countries, public investment on infrastructures, through EU structural programs, has been
the instrument of choice to induce real convergence of the domestic economy to the EU standards of living.
Furthermore, in the near future, the eastward expansion of the EU will bring into the fold countries with
similar problems. For these Eastern European countries, economy recovery seems to depend, in large scale,
on the reconstruction of obsolete infrastructures. For these countries joining the EU and, thereby, embarking
in large public infrastructure projects seems to be the expected vehicle for vanquishing their relative
backwardness.
The empirical evaluation of the effects of public capital formation on private output was brought to
the limelight by the work of Aschauer (1989a, 1989b). Using a single-equation static production function
approach based on aggregate measures of public capital, Aschauer (1989a, 1989b) suggests that public capital
has been a powerful engine for growth in the United States. In fact, his results suggest that public investment
would pay for itself close to three times in the form of additional tax revenues over the duration of the public
capital assets [see Reich (1991)]. Subsequent analysis applying the same methodology to regional and sectorspecific data in the United States as well as international data, however, failed to replicate such large effects.
Indeed, it often even failed to find meaningful positive results [see Gramlich (1994) and Munnell (1992) for
3
detailed surveys of the literature and Hulten and Schwab (1993) for a detailed presentation on the
infrastructure debate].
The work of Aschauer inspired an important body of literature on the impact of infrastructure
development for other countries. This includes contributions that are country-specific and others that have a
multi-country focus. In the first case one could mention, for example, the work of Otto and Voss (1996) for
Australia, Seitz (1994) for Germany, Sturm and de Haan (1995) for Holland, Merriman (1990) for Japan,
Shah (1992) for Mexico, Pereira and Roca (1999) for Spain, Berndt and Hansson (1992) for Sweden, and
Lynde and Richmond (1993) for the UK. In the second case one could mention, for example, the work of
Aschauer (1989c), Evans and Karras (1993), Ford and Poret (1991), and Mittnik and Newman (1998), all
focusing on developed OECD countries. The magnitude and significance of the empirical results varies
greatly among countries. Furthermore, international comparisons are rendered very difficult by the use in the
literature of different measures of public capital, different levels of aggregation, and different methodologies.
The approach used in Aschauer (1989a, 1989b) and much of the literature that followed focuses on
measuring the effects of public investment on private output using a single equation, static production function
approach. In this approach, private output is regressed on public capital and private inputs – employment and
capital. This approach has been criticized on econometric grounds. It has been observed that the estimation
of static, univariate production functions in levels (or log-levels) is based on non-stationary variables.
Therefore, OLS estimates are spurious in the absence of cointegration. Moreover, OLS estimates suffer from
simultaneity bias. Even if this bias is corrected, conclusions about causality still cannot be drawn. [See
Jorgenson (1991) and Munnell (1992) for comprehensive discussion of these econometric problems.]
In this paper, we follow Pereira (2000) and adopt a vector auto-regressive/error correction
mechanism approach. This multivariate time series approach allows us to address the aforementioned
econometric criticisms in a rigorous and comprehensive manner. It also brings a more precise conceptual
focus to the debate about whether or not public capital is productive. In fact, the static single-equation
framework, so often used in the literature, excludes the presence of feedbacks, in particular dynamic
feedbacks, among the relevant variables. This exclusion is of paramount importance for it is likely that
4
feedbacks exist. If they do, a zero elasticity of private output with respect to public capital, as obtained from a
single-equation static production function approach, is neither a necessary nor a sufficient condition for public
investment to be ineffective in influencing output.
Dynamic feedbacks are essential to a conceptual understanding of the relationship between public
investment and aggregate economic performance. Indeed, public investment affects output directly as an
additional input in the production function. Moreover, as a positive externality to aggregate production,
public investment should, ceteris paribus, lead to higher aggregate production. Public investment also affects
aggregate production indirectly via its effects on the use of private inputs, capital and labor. It is conceivable
that a greater availability of public capital could reduce the demand for private inputs (a substitution effect).
Higher availability of public capital, however, also increases the marginal productivity of private inputs. This
lowers the marginal costs of production, thereby potentially increasing the level of aggregate production (a
scale effect).
In turn, the evolution of private inputs and aggregate output can conceivably affect the evolution of
public investment. Indeed, increasing aggregate output provides the government with a growing tax base and
the potential for greater public investment. Furthermore, declining employment has often led to short-term
policy packages that involve increased public investment. There is, therefore, a real possibility that reverse
causality exists. By this we mean that it is possible that the evolution of aggregate output and private inputs
may be leading the evolution of public investment.
This paper is organized as follows. In Section II, we present the data set used in our analysis. In
particular, we present in some detail the new public investment data for Portugal (see also the Appendix). We
also report preliminary empirical results including univariate and cointegration analysis and report on the
specification of the vector auto-regressive/error correction mechanism models. In Section III, we introduce
and discuss some methodological issues in the identification and measurement of the effects of innovations in
public investment. In Section IV, we analyze the effects on economic performance - output, employment, and
private investment - of aggregate and disaggregate measures of public investment through the use of
orthogonalized impulse response functions. Finally, in Section V, we provide a summary and some
concluding remarks.
5
II. Data and Preliminary Empirical Results
A.
Data: sources and description
We use annual data for the period 1976 to 1998. We consider output (gdp), employment (emp),
private investment (inv), in addition to public investment in transportation infrastructures (pinv). The data on
output, employment, and private investment in presented in Table 1. This data was obtained from the Bank of
Portugal/Banco de Portugal (1997), Commission of the European Communities (1999), and Ministry of
Finance/Ministério das Finanças (2000). Output and private investment are measured in millions of constant
1995 Portuguese escudos while employment is measured in full-time equivalent employees.
The data for public investment in transportation infrastructures (pinv) is obtained from Pereira and
Andraz (2001). This database is the result of a long and meticulous investigation, sponsored by the
Portuguese Ministry of Planning. This database includes data on public investment, both at current and
constant 1995 prices deflated by the GDP as well as the private investment deflators. It includes public
investment in national roads, municipal roads, highways, ports, airports, and railways. It covers the period
from 1974 to 1998, despite some failures of information regarding the two first years, due to lost data at the
source. Since this database has not been published before and is used in this article for the first time, it is
provided here in Table 2 and we discuss some of its main features below. For the same reason we also
included in an Appendix to this paper the executive summary of Pereira and Andraz (2001). All of the data is
in 1995 Portuguese escudos deflated using the GDP price deflator. The use of the private investment deflator
would lead to only marginal changes in the empirical results in this paper.
To talk about the main features of the public investment data in Portugal one has immediately to
recognize the existence in the second half of the sample period of EU sponsored structural transfer programs
in the form of Community Support Frameworks for Portugal. The first Community Support Framework
6
program covered the period from 1989 to 1993 and the second covered the period from 1994 to 1999.
Therefore, our sample includes 13 years prior the programs and 10 years of with the programs.
In what follows we consider an aggregate measure of public investment in transportation
infrastructures, as well as six disaggregated measures pertaining to public investment on roads, ports, airports
and railways. We present the evolution of each type of investment as a percentage of the GDP and as a
percentage of private investment in Tables 3 and 4, respectfully, and in Figures 1-7. We present the evolution
of the composition of public investment in transportation infrastructures in Table 5 and Figure 8.
The first type of public investment (pinv1) is core infrastructure investment in national roads. It
averages 0.48% of the GDP for the sample period. It experiences a strongly increasing trend during the
sample period, from 0.34% of the GDP in the early years of the sample to 0.76% by the end of the sample
period. The second type of public investment (pinv2) is core infrastructure investment in municipal roads. It
averages 0.40% of the GDP for the sample period and shows less of a variation in that it averages 0.35% in
the first part of the sample and 0.45% in the second. The third type of public investment (pinv3) is core
infrastructure investment in highways. It represents an average of 0.21% of the GDP over the sample period,
although the average in the early years is just 0.13% and in the second part of the sample is 0.32%. The fourth
type of public investment (pinv4) is core infrastructure investment in ports. It represents on average 0.12% of
the GDP and has experienced a decline from 0.15% in the 1970s and 80s to about 0.08% in the last decade.
The fifth type of public investment (pinv5) is infrastructure investment in airports, and has remained stable
over the sample period at about 0.05% of the GDP. Finally, the sixth type (pinv6) is core infrastructure
investment in railways. It averages 0.29% of the GDP for the period and it only shows an upward trend in the
last few years of the sample.
Overall, aggregate public investment (pinv) averages 1.55% of the GDP for the sample period. It
changes, however, from an average of 1.24% for 1976-88 to an average of 1.96% for 1989-1998. The data
suggests that the increase through the 1990s in the overall figures is due mostly to increases in public
investment in national roads (pinv1) and highways (pinv3) and, more recently, in railroad investment (pinv6).
All of the considerations above suggest that the data fully reflect the conventional wisdom that the
EU structural transfer programs brought a greater dynamism to the public investment in infrastructures. They
7
are also very informative about the effects of the EU Community Support Frameworks in terms of the
composition of public investment in transportation infrastructures. In fact, core investment in national roads
(pinv1) is one of the greatest beneficiaries of these programs. Its share on total public investment increased
from 27.4% in the 1980s to 33.7% in the 1990s. Core investment in highways (pinv3) also increased its share,
from 10.1% to 16.5%. Core investment in railways (pinv6), was also positively affected by EU programs.
During the period between 1989 and 1998, it represents 19.3% of total public investment after having
accounted for about 17% until 1988. On the other side of the spectrum are the other types of public investment
whose shares decreased during the period covered by EU programs. The share of public investment on
municipal roads (pinv2) declined from 29.2% to 23.3%, while the share of investment on airports (pinv5)
declined from 3.7% to 2.9%. However, the greater losses occurred in public investment on ports (pinv4).
From a share of 12.1% in the period before 1989, it represents only 2.9% of total public investment during the
1990s.
Besides the changes in magnitude and composition of public investment before and after the
Community Support Frameworks, it is also possible to detect some changes from the first program (1989-93)
and second (1994-98). The shares of public investment on national roads (pinv1) and on railways (pinv6)
show an increasing trend during the two structural programs. Their shares to total public investment during the
second program are higher than the average share for the 1990s. The share of public investment on highways
(pinv3), decreased from 17.4% to 15.7% to total public investment during the second program. Public
investment on municipal roads (pinv2) shows a continuous decreasing pattern during the 1990, from 29.2%
before the structural programs, to 25.9% and 20.6% during the first and the second programs, respectively.
Public investment on ports (pinv4), whose share to total was 12.1% in the period until 1988, suffered a sharp
decline during the first program to 5.3%. During the second program, its share still decreased to 3.5%. Finally,
the share of public investment on airports (pinv5) decreased to 3.4% during the first program, and to 2.4%
thereafter.
We conclude this discussion of the public investment data with some brief international comparisons.
International data comparisons are very difficult. This is mostly due to the fact that the definition of the
public investment data and its scope vary greatly across countries. Furthermore, detailed disaggregated public
8
investment data sets are not readily available for most countries. Despite these cautionary words we provide a
tentative comparison of aggregate public investment in transportation infrastructures in Portugal, Spain and
the United States. For the Spanish data sources and specific definitions see Pereira and Roca (1999) and for
the US case see Pereira (2000). We present the evolution of public investment in transportation infrastructures
as a percentage of the GDP for these three countries in Figure 9.
There are two aspects that are worth mentioning. First, aggregate public investment in transportation
infrastructures in Portugal is of the same order of magnitude as in Spain. In both countries public investment
in transportation infrastructures tends to be somewhat above the levels for the US, in particular after the late
1980s. This is an obvious implication of the EU structural programs, which have been in effect for both
countries since 1989. Second, the upward trend that can be detected after the late 1980s in both Portugal and
Spain is much less pronounced in the Spanish case after 1993. This can also be explained by the
characteristics of the EU structural programs. Indeed, the EU structural programs for Spain became less
important after 1994, with the inception of the second Community Support Framework.
B.
Univariate and cointegration analysis
We start by using the Augmented Dickey-Fuller (ADF) t-test to test the null hypothesis of a unit root
in the different variables. We use the Bayesian Information Criterion (BIC) to determine the optimal number
of lagged differences to be included in the regressions, and we include deterministic components, a constant
and/or a trend, in the regressions if they are statistically significant.
The results of the ADF t-tests applied to the different variables in log-levels, are presented in the top
part of Table 6. In all the cases, the t-statistics are lower, in absolute levels, than the 5% critical values.
Therefore, the ADF tests cannot reject the null hypothesis of a unit root in these variables. In turn, the results
of ADF t-tests applied to the first differences of the log-levels, i.e., the growth rates of the original variables,
are presented in bottom part of Table 6. All critical values are greater, in absolute value, than the 5% critical
value. Therefore, we can reject the null hypothesis of unit roots in the growth rates of the variables. We take
this evidence as an indication that stationarity in first differences is a good approximation for all the time
series under consideration.
9
We also test the null hypothesis of a unit root in the different variables using the Phillips-Perron test,
which takes into consideration the possible existence of structural breaks in the evolution of the variables. This
is an important step since due to the different EU structural programs structural breaks are likely to exist. We
follow the same strategy as above in the determination of optimal lags and deterministic components in the
tests. The test results are reported in Table 7. The results from the Phillips-Perron unit roots tests completely
confirm the previous unit root test results. Again the strong evidence is that stationarity in first differences is a
good approximation for all the time series under consideration.
It should be pointed out that this empirical evidence is consistent with the conventional wisdom in the
macroeconomics literature that aggregate public investment, output, employment, and private investment are
stationary in first differences. Although our public investment series is more disaggregated, the same pattern
of stationarity in first differences is not surprising.
We now test for cointegration among output, employment, aggregated private investment, and
aggregated public investment as well as each one of the six public investment variables. We use the standard
Engle-Granger approach to test for cointegration. We have chosen this procedure over the often-used
Johansen approach for two reasons. First, since we do not have any priors that suggest the possible existence
of more than one cointegration relationship, the Johansen approach is not strictly necessary. More
importantly, however, for small samples, Johansen's tests are known to induce strong bias in favor of finding
cointegration when it does not exist. [See, for example, Gonzalo and Lee (1998).] Therefore, our relatively
small sample size suggests that the standard Engle-Granger approach will lead to more accurate results.
Following the standard Engle-Granger approach, we perform four tests in each case. This is because
it is possible that one of the variables will enter the cointegrating relationship with a statistically insignificant
coefficient. We do not know, a priori, whether or not this will happen. If it does happen, however, a test that
uses such a variable as the endogenous variable will not pick up the cointegration. Therefore, a different
variable is endogenous in each of the four tests. We apply the ADF t-test to the residuals from the regressions
of each variable on the remaining variables. In all of the tests, the optimal lag structure is chosen using the
BIC, and a deterministic component is included if it is statistically significant.
The results of the cointegration tests at the aggregate level are reported on the top part of Table 8.
10
The value of the t-statistics is lower, in absolute value, than the 5% critical values for at least three of the four
cases considered. Moreover, all the test statistics are lower, in absolute value, than the 1% critical values.
Thus, the ADF tests cannot reject the null hypothesis of a random walk, and we cannot reject that the variables
are not co-integrated at this aggregated level.
Cointegration tests were also performed with aggregate output, employment and private investment,
together with each of the six different types of public investment. Results are also reported in Table 8. For all
six public investment variables, the value of the t-statistics is lower, in absolute value, than the 5% critical
values for all but four of the twenty-four cases considered. We take this as strong evidence that, consistently
with the results at the aggregate level, the variables are not cointegrated at the more disaggregated level.
C.
VAR specifications and estimates
We have now determined that all of the variables have the same order of integration and, in
particular, that they are stationary of first order. We have also determined that the variables do not seem to be
cointegrated, either at the aggregate level or at the more disaggregated level. Accordingly, we follow the
standard procedure in the literature and determine the specifications of the VAR models using growth rates of
the original variables (denoted by ggdp, gemp, ginv, gpinv, etc).
We estimate seven VAR models. All VAR models include aggregate output, employment, and
private investment. In addition, each of the seven VAR models includes a different public investment variable
- one for aggregate public investment and one for each of the six different types of public investment. This
means that, consistently with our conceptual arguments, the public investment variables are endogenous
variables throughout the estimation procedure. For the sake of brevity, the details on the model selection for
the different VAR models are not reported here. They are available from the authors upon request.
The specifications of the different VAR models are determined using the BIC. The test results are
reported in Table 9. The VAR specification has two dimensions, which were determined jointly - the
specification of the deterministic components and the consideration of the possibility of structural breaks. In
all cases a first order specification were selected. A higher order was not considered due to relative small size
of sample. The BIC selects a specification with constant and trend for the disaggregated models for national
11
roads, municipal roads, highways, and ports. For the aggregate model, as well as the disaggregated models for
airports and for railways, the BIC selects a specification with only a constant.
In order to consider the possible structural changes due to the two Community Support Frameworks,
different VAR specifications were considered. One could possibly distinguish three periods in which there
might have been structural changes: the period before 1989, the period of first program, i.e., 1989-93 and the
period of the second program, i.e., 1994-98. Therefore, we consider three alternatives in terms of the VAR
specification. The first is the case of no structural break/no dummies. The second is the case of one structural
break/one dummy distinguishing the periods before and after the EU structural programs. Finally, we consider
the possibility of two structural breaks/two dummies reflecting the possibility of the three different periods
mentioned above. We find that the BIC criterion leads to the selection of VAR with two structural breaks/two
dummies for aggregate public investment as well as for each one of the six types of public investment. This
suggests that in addition to considering the differences before and after the EU structural programs, there are
also important changes associated with each of the EU structural programs.
III.
Identifying and Measuring the Effects of Innovations in Public Investment
We use the impulse-response functions associated with the estimated VAR models to examine the
effects of the different types of public investment on the performance of output, employment and private
investment variables. In this context our methodology allows dynamic feedbacks among the different
variables to play a critical role. This is true in both the identification of innovations in the public investment
variables and the measurement of the effects of such innovations.
A.
Identifying innovations in the public investment variables
While the public investment variables are endogenous in our econometric framework, the key
methodological issue for the determination of the effects of public investment on the other variables is the
identification of innovations in the public investment variables that are truly exogenous. This means that we
12
need to identify shocks to public investment variables that are not contemporaneously correlated with shocks
in the remaining variables. These exogenous shocks are not subject to the reverse causation problem. In
dealing with this issue we draw from the approach typically followed in the literature on the effects of
monetary policy on the economy [see, for example, Christiano, Eichenbaum and Evans (1996, 1998), and
Rudebush (1998).]
Ideally, the identification of shocks to public investment which are uncorrelated with shocks in other
variables would result from knowing what fraction of the government appropriations in each period is due to
purely non-economic reasons. The econometric counterpart to this idea is to imagine a government policy
function which relates the rate of growth of public investment to the information in the relevant government
information set; in our case, the past and current observations of the growth rates of the output, employment
and private investment variables. The residuals from this policy function reflect the unexpected component to
the evolution of public investment and are uncorrelated with other innovations.
In the central case, we assume that the relevant information set for the government policy function
includes past values but not current values of the other variables. This is equivalent in the context of the
standard Choleski decomposition to assuming that innovations in public investment lead innovations in the
other variables. This means that we allow innovations in public investment to affect the other variables
contemporaneously, but not the reverse.
We have two reasons for making this our central case. First, it seems reasonable to believe that the
private sector reacts within a year to innovations in public investment decisions. Second, it also seems
reasonable to assume that the public sector is unable to adjust public investment decisions to innovations in
the private-sector variables within a year. This is due to the time lags involved in information gathering and
public decision-making. Nevertheless, to determine the robustness of our central case results, we also consider
all the possible alternatives in terms of the definition of which observations of the private sector variables are
included in the government information set. This is equivalent to considering all the possible orderings of the
variables within the Choleski decomposition framework. We report the corresponding range of results for the
variance decomposition in Table 10 and for the impact indicators in Table 11.
It should be pointed out that the sensitivity analysis efforts could conceivably be generalized in two
13
different directions. First, we could consider the effects of innovations in the private sector variables, for
example a supply shock, under our current sensitivity analysis framework. To do so, however, would require
a great deal of assumptions as to the ordering of the private sector variables. Our approach has the advantage
of providing a measure of the effects of innovations in public investment variables on private sector variables
that is independent of the ordering of the private sector variables. We can, therefore, remain agnostic about
the issue of the order of these variables. Second, we could generalize the sensitivity analysis framework to
consider non-recursive or signal extraction schemes. This would reflect, however, an econometric more than
an economic concern. It would only be justified if we had less strong priors about what the central case
should be and it would entail alternatives of less clear economic interpretation. Because of these reasons we
have not pursued either path in this paper.
B.
Policy functions
The policy functions for aggregate public investment as well as the different types of public
investment are reported in Table 12. These policy functions relate the evolution of the public investment
variables to the evolution of the private sector variables lagged one year, according to the selected VAR
specification. For the aggregate model, there is no feedback from the other variables to public investment.
This means that public investment is truly an exogenous variable.
It is interesting to note that the exogeneity of public investment in Portugal is in contrast with the
findings for the US, for example. In fact, Pereira (2000) shows that changes in public investment in the US
are positively correlated with lagged changes in output and negatively correlated with lagged changes in
employment. Therefore, changes in private-sector variables affect the evolution of public investment in the
US, which is not an exogenous variable. The exogeneity of public investment decisions in Portugal, however,
is easily explained by the fact that for long public investment decisions have been closely related with the
Portuguese participation in the EU. This is particularly true after 1989, when the bulk of the public
investment in transportation infrastructures in Portugal has been conducted under the two Community Support
Frameworks. These programs are typically negotiated between the recipient economies and the EU, focusing
on long-term goals and deliberately avoiding short-term considerations.
14
It should be pointed out that while public investment seems to be an exogenous variable at the
aggregate level, the aggregate results hide some important effects on the evolution of different types of public
investment. This means that although the magnitude of public investment seems to be truly exogenous there
may be some effects from the economy on the composition of public investment.
In fact, the policy functions suggest that changes in the evolution of public investment in national
roads (pinv1) respond positively to changes in output while the evolution of public investment in municipal
roads (pinv2) depends positively on the evolution of private investment. In turn, the evolution of investment in
highways (pinv3) depends positively on the evolution of employment and negatively on the evolution of
private investment. Public investment in ports (pinv4) depends positively on lagged changes in output and
negatively on lagged changes in private investment. Finally, the evolution of public investment in airports
(pinv5) and in railroads (pinv6) does not seem to respond to lagged economic performance.
C.
The impulse-response functions
We consider the effects of one-percentage point, one-time random shocks in the rates of growth of
the different types of public investment on output, employment, and private investment. We expect these
temporary shocks in the growth rates of the different types of public investment to have temporary effects on
the growth rates of the other variables. They will, however, have permanent effects on the levels of those
variables. The accumulated impulse response functions are reported in Figures 10-16.
There are a few interesting points worth mentioning in terms of these accumulated impulse-response
functions. Let us start by acknowledging that all accumulated impulse-response functions converge within
approximately a five-year period. This is not inconsistent with the idea that public investment takes time to
build before it really impacts the private sector performance. This is because our measures of public
investment are aggregate measures, which are made of spending from a series of overlapping public
investment projects. This being the case, in any given year a substantial part of the observed public
investment corresponds to projects that have been concluded that year.
It should also be noted that the convergence path of the private sector variables is not only relatively
fast but also very smooth. In turn, the convergence path of the public investment variables, although fast, is
15
less smooth in the early years. This pattern can easily be understood if one considers that the initial exogenous
shock to public investment variables is followed by an endogenous adjustment in public investment in
response to the private sector variables. This endogenous adjustment is dictated in the context of the VAR
model by the policy functions presented in Table 12 and discussed above.
These policy functions suggest a negative recursive pattern in the evolution of public investment in
addition to their response to private sector variables. This negative recursive pattern dominates in the early
years while the effects on the private sector variables are relatively small. This explains the dip in the
impulse-response function in the early years. In later years, however, the positive feed back from the
evolution of the private sector variables seems to dominate, although in some cases it is not strong enough to
bring the accumulated long-term change in public investment to its initial level on impact. Hence, for
aggregate public investment (pinv), for example, the long-term change in public investment associated with a
1.0 percentage point change on impact is 1.2 approximately, while for national roads (pinv1) is about 1.0, and
for municipal roads (pinv2) is about 0.8.
D.
Measuring the effects of innovations in public investment variables
In this paper we estimate the long-term accumulated elasticities of the different variables with respect
to each type of public investment. Long-term is defined as the time horizon over which the growth effects of
innovations disappear, i.e., the accumulated impulse-response functions converge. These elasticities represent
the total percentage point changes in the different variables for each long-term accumulated percentage point
change in public investment once all the dynamic feedback effects among the different variables have been
considered.
We report the long-term accumulated marginal productivity of public investment in terms of the other
variables in Tables 13 to 15. These figures measure the change, in million euros, in output and private
investment for every million euros accumulated change in public investment. In Table 14, we report the effects
in terms of the number of jobs created in the long-term per one million euros in public investment.
We obtain the marginal product of private investment reported in Table 15, by multiplying the output
to public investment ratio for the last ten years by the elasticity of output with respect to public investment.
16
The choice of the output to public investment ratio for the last ten years is designed to reflect the relative
scarcity of public investment of the different types. We consider the relative scarcity at the margin of the
sample period without letting these ratios be overly affected by business cycle factors or by the different
priorities established by EU structural programs.
It should be noted that we use the term marginal product in a way that departs from the conventional
definition of the word. This is because these elasticities and marginal products are not based on ceteris paribus
assumptions. In this paper, the term marginal product includes all of the dynamic feedbacks among the
variables. Therefore, the marginal product that we calculate is a total marginal product. That is, it measures
both the direct effects of public investment on output, and the indirect effects of public investment on output
through changes in the evolution of inputs. Of course, this is the relevant concept from the standpoint of
policy making.
Finally, the annual rates of return, also reported on Table 15, are calculated from the marginal
product figures by assuming a life horizon of twenty years for all types of public capital assets. That is, the
rate of return applied to one euro over a twenty-year period yields the value of the accumulated marginal
product. These rates of return are adjusted to accommodate for a public capital depreciation rate of 5%,
which is implicit in the life horizon of twenty years for the public capital assets.
IV. On the Economic Effects of Public Investment
A.
Aggregate effects of public investment in transportation infrastructure
The effects on employment and private investment of public investment at the aggregate level are
reported on the top part of Tables 13 and 14. The results from the impulse response analysis at the aggregate
level suggest that in Portugal, public investment in transportation infrastructure crowds in both private
investment and employment. Indeed, when we estimate the effects of shocks to aggregate public investment in
transportation infrastructures on the evolution of the other variables, we find that the elasticity of private
investment with respect to aggregate public investment is 0.639, which corresponds to a marginal product of
17
8.1. This means that at the aggregate level, public investment crowds in private investment and that one euro
of additional public investment will induce, in the long-term, an accumulated total of 8.1 euros of private
investment. In turn, the elasticity of employment with respect to aggregate public investment is 0.079. This
figure suggests that 230 additional jobs will be created in the long-term for each additional one million euros
in public investment in transportation infrastructures.
In turn, the long-term effects of innovations in investment in transportation infrastructures on output
are reported on the top part of Table 15. We find that aggregate public investment has a positive effect on
output with an elasticity of 0.183, which corresponds to a marginal product of 9.5. This implies that the
increase of one euro in public investment leads to a total accumulated increase of 9.5 euros in output.
One possible way of interpreting the marginal product figures is by calculating the corresponding
average rate of return. The estimated annual rate of return over a twenty-year period of public investment in
transportation infrastructures is 15.9%. This figure suggests that the rate of return of public investment in
transportation infrastructures is well above the range one would expect for the rate of return on private
investment. From this perspective, the reliance on public investment in transportation infrastructures as the
cornerstone of a development strategy in Portugal seems to have been justified.
Another possible way of interpreting this figure is by calculating the value of the tax revenues
generated by this increase in output. Since tax revenues tend to hover around the 35% of the GDP, then the
marginal product of public investment in transportation infrastructures suggests that over the life expectancy
of the public capital assets, the public sector would collect 3.3 euros. Therefore, the public sector collects an
additional 3.3 euros in tax revenues for each euro spent in public infrastructure. According to this evidence,
the public investment assets in transportation infrastructures pay for themselves over their life span and still
generate additional funds, which can be used for other public activities.
B.
Effects of public investment in different types of transportation infrastructure
In the discussion above, we have established empirically that public investment in transportation
infrastructure makes a positive and significant contribution to private-sector performance. We are ready to
determine which types of public investment are the most productive. The positive crowding in effects of
18
public investment in transportation infrastructures on private investment and employment observed at the
aggregate level are also present at the disaggregated level. All types of public infrastructure in transportation
affect private investment and employment positively in the long-term. Not surprisingly, the same pattern can
be found in terms of the effects on output.
The effects of public investment in transportation infrastructures on private investment are reported
in Table 13. In terms of the effects of public investment on private investment, the strongest effect comes
from public investment in national roads (pinv1) with an elasticity of 0.766. It is followed, by the investments
in municipal roads (pinv2), ports (pinv4), and railways (pinv6) with elasticities of 0.254, 0.281 and 0.264,
respectively. Finally, public investment in highways (pinv3) and in airports (pinv5) display the lowest
elasticities, 0.110 and 0.079, respectively.
In terms of marginal productivities, a better measure of relative scarcity, the highest marginal effects
on private investment come from public investment in ports (pinv4) and airports (pinv5) with marginal
products of 84.4 and 39.1 respectively. The marginal products of public investment in national roads (pinv1),
municipal roads (pinv2), and railroads (pinv6) are still relatively large – 29.6, 14.1, and 18.8, respectively.
The lowest effects on private investment come from public investment in highways (pinv3) with a marginal
product of 9.2.
The effects of public investment in transportation infrastructures on employment are reported in
Table 14. The strongest effect comes now from shocks to ports (pinv4) with an elasticity of 0.070, followed by
municipal roads (pinv2), with an elasticity of 0.054, national roads (pinv1), with an elasticity of 0.045. In
turn, the elasticities of public investment in highways (pinv3) and railways (pinv6) are substantially smaller,
0.009 and 0.012, respectively. Finally, the effect on employment of public investment in airports (pinv5) is
only marginally different from zero.
In terms of job creation, one million euros invested in ports (pinv4) will create, in the long-term,
about 4800 new jobs. This number reduces sharply to 692, 404, 204, and 164 new jobs per million euros
invested in municipal roads (pinv2), national roads (pinv1), railways (pinv6), and highways (pinv3)
respectively. Finally, public investment in airports (pinv5) actually eliminates about 500 jobs per million
euros.
19
The effects of public investment in transportation infrastructures on output are reported in Table 15.
The effects of shocks to the different public investment variables on output are all positive. In terms of the
long-term accumulated elasticities, the strongest effect comes from shocks to public investment in national
roads (pinv1) with an elasticity of 0.198. This is followed by the effect of shocks in public investment in
municipal roads (pinv2), with an elasticity of 0.098, in ports (pinv4), with an elasticity of 0.087, and railways
(pinv6), with an elasticity of 0.062. In turn, the elasticities of output with respect to public investment in
highways (pinv3) and airports (pinv5) are the smallest, respectively 0.024 and 0.009.
Let us now consider the marginal product figures. These figures are a better measure of the relative
effects of different types of public investment. This is because they reflect the relative scarcity of the different
types of public investment at the margin of the sample period. The marginal product figures suggest that all
types of public investment are productive. Although there is a wide range of effects, four of the six types of
transportation infrastructure have marginal products within a relatively small range, between 18.5 and 31.4.
This is the case of public investment in national roads (pinv1), municipal roads (pinv2), airports (pinv5) and
railroads (pinv6), with marginal products of 31.4, 21.3, 19.2, and 18.5, respectively. The two extremes are
given by public investment in highways (pinv3) with a marginal product of just 8.2 and public investment in
ports (pinv4) with a marginal product of 107.1.
Another way of interpreting these results is by considering the rates of return on the different types of
public investment. Again, all rates of return for all different types of public investment in infrastructures are
above the expected ranges for private investment. Over a twenty-year period, the average rate of return to
public investment in ports (pinv4), is 30.8%, and is the highest. It is closely followed by the rate of return to
public investment in national roads (pinv1) of 23.0%, municipal roads (pinv2) of 20.9%, airports (pinv5) of
20.0%, and railroads (pinv6) of 19.7%. The lowest rate of return, although still high, is for public investment
in highways (pinv3) with 15.0%.
It is important to highlight the importance of considering both the direct and the indirect effects of
innovations in public investment. The explicit consideration of the indirect effects of public investment on
private investment and employment allows us to highlight the mechanisms through which the different types of
public investment tend to affect output. Indeed, the strong effects of public investment in ports (pinv4) on
20
output, seems to be related to strong effects on both employment and private investment. The converse is true
for public investment in highways (pinv3) in which case the less strong effects seem to be related to less strong
effects on also both private investment and employment. In turn, the effects on output of public investment in
municipal roads seems to be due mostly to the important effects on employment while the effects of public
investment in airports (pinv5) seem to be mostly related to the effects on private investment.
C.
Comparison with the international evidence
The comparison of the results in this paper with the international evidence available in the literature
is not easy. This is primarily because the international literature has used a variety of econometric techniques,
which makes similar terms, like elasticity or marginal product not always comparable with the way such terms
are used in this paper. Also, most of the literature on the effects of public infrastructures considers public
investment as an exogenous variable and focuses on the effects of public investment on private output and is
not designed to address the impact on private inputs. Furthermore, the definitions of public investment used in
the literature vary wildly.
Although comparisons are difficult they are not impossible. The results in this paper are most
directly comparable with the results in Pereira and Roca (1999) for Spain and in Pereira (2000) for the US.
Pereira and Roca (1999) consider for Spain the effects of public capital in transportation infrastructures. The
empirical results suggest a marginal product of private investment with respect to public investment of 10.2
and that one million euros in public investment create 129 jobs in the long-term. Moreover, the results indicate
that the marginal product of public investment in Spain is 5.5. This corresponds to a rate of return of 8.9%.
Accordingly, the results obtained in this paper for Portugal, 230 new jobs created per million euros in public
investment and a rate of return of 15.9%, tend to be higher than the ones for Spain.
In turn, Pereira (2000) finds that public investment, although under a much broader definition,
crowds in private investment with a marginal product of 0.8 while it seems to have a negligible effect on
private employment. The results in this paper for Portugal show much larger figures for the marginal effects
of public investment in transportation infrastructures on private investment – about ten times, while the effects
on employment in Portugal are substantial - 230 jobs per one million euros in public investment. More
21
importantly, Pereira (2000) suggests that the marginal product of public investment in the US is 4.5. This
corresponds to a rate of return of 7.8%, compared to a rate of return of 15.9% in the Portuguese case. Again,
the results in this paper tend to be substantially higher than the results for the US. This is understandable
given the relatively greater scarcity of public infrastructures in the Portuguese economy.
International comparisons in terms of the disaggregated effects of different types of public investment
are even more difficult. Again, probably the closest comparisons can be made with Pereira (2000). In Pereira
(2000) there is a core infrastructure variable, which represents highways and streets and that closely resembles
the aggregate of national roads, municipal roads, and highways. Another variable in Pereira (2000), core
infrastructure in ports, airports, etc, seems to be close to the aggregate of ports and airports in this paper. The
results indicate that the marginal products of these two types of public investment are 1.97 and 19.79,
respectively. The correspondent rates of return are 3.4% and 16.1%, respectively. In this paper, the range of
rates of return is from 15.0% to 23.0% for public investment in roads and highways, and 30.8% and 20.0% for
public investment in ports and airports, respectively. Again, the figures for Portugal tend to be substantially
higher than the ones for the US.
An important feature of the empirical results in this paper is that in Portugal public investment in
transportation infrastructures would more than pay for itself in the form of added tax revenues over the life
span of the public investment assets. This is reminiscent of the supply-side Laffer-curve effect found for the
United States by the early literature. Indeed, the seminal contribution of Aschauer (1989a) has been
interpreted as suggesting that [see, for example, Reich (1991)]. This result was disputed by subsequent
research for the United States case. For example, Pereira (2000) suggests that the marginal product of public
investment would just pay for itself over time. Furthermore, Pereira and Roca (1999) show that the same is
true for Spain while the results in Mittnik and Newman (1998) for Canada, France, Germany, Japan, The
Netherlands, and the United Kingdom, in a time series context not incompatible with the approach in this
paper, seem to imply the same. Interestingly enough, however, the same type of result seems to resurface in
the case of Portugal. This leaves open the question as to whether a supply-side Laffer-curve effect while not
present in more developed economies could be a fixture of less developed countries.
22
V. Summary and Concluding Remarks
This paper analyzes empirically the effects of public investment in transportation infrastructure on
economic performance in Portugal. To do so, we use a new data set on public investment in transportation
infrastructures in Portugal for the period 1976-98, recently published by Pereira and Andraz (2001). We
follow a VAR approach, which is consistent with the argument that the analysis of the effects of public
investment on output, employment and private investment variables requires the consideration of dynamic
feedback effects among the different variables.
We can summarize the empirical results as follows. We find that in the long-term, aggregate public
investment in transportation infrastructures crowds in private investment as well as employment. More
importantly, we find that it has a positive effect on output. Indeed, we estimate that one euro invested in
public investment increases output in the long-term by 9.5 euros. This figure suggests that public investment
pays for itself 3.3 times in the form of tax revenues over the life span of the public capital asset. Furthermore,
the marginal product figure corresponds to a rate of return of 15.9%. This rate of return is clearly higher than
the rate of return expected on private investment activities.
The importance of the effects of public investment in transportation infrastructures at the aggregate
level opens the door to the next stage of our analysis: the study of the effects of different types of public
investment on economic performance. Consistent with the aggregate results, we find that all types of public
investment crowd in the other variables. Nevertheless, a close look at the effects of different types of public
investment on the remaining variables suggests that the disaggregation of public investment is very
informative, since it shows which types of public investments are the most productive. In terms of marginal
productivities, the highest effects on private investment come from public investment in ports, airports and
national roads. In terms of job creation, the highest effects come from public investment in ports, municipal
roads, and national roads. Finally, in terms of the effects of output the largest effects come from investment in
ports followed closely by national roads, municipal roads, airports, and railroads.
The results in this paper are very important and timely from a public policy perspective in Portugal.
23
From a retroactive perspective, the empirical evidence suggests strongly that public investment in
transportation infrastructures has been a powerful instrument to promote long-term growth in Portugal.
Moreover, it suggests that the strategy followed by the Portuguese authorities of investing in public
infrastructures has been justified both from a long-term development perspective as well as from a public
budgetary perspective.
More importantly, from a prospective perspective, the results in this paper provide broad guidelines
for the country's future development strategies. This is very important due to the still relative backwardness
now. As a matter of fact, Portuguese GDP is still at about 75% of EU average while sources of outside
financing are being reduced and the country faces a great budgetary restraint in the context of EMU. This
requires greater attention to relative benefits and much more fine-tuned development policies. It is also
important to highlight the fact that given current budgetary constraints in the context of the Stability and
Growth Programs, the tendency for achieving budgetary consolidation through reduction in public investment
is a mistake from the standpoint of long-term growth. It is also a mistake from the standpoint of long-term
budgetary situation.
Although the results in this paper are important from the perspective of policy making in Portugal, its
interest is far from parochial. In fact, there is a number of Eastern European waiting to join the EU. These
countries have levels of development and infrastructure scarcities that are not unlike the Portuguese case by
the end of 1980s. Furthermore, there are already structural transfer programs in place to smooth the transition
of these countries into the EU and they are expected to benefit from large EU structural funds upon accession,
much like Greece, Ireland, Portugal, and Spain currently do. From this paper we learn that the general
strategy of investing in public infrastructure may be very effective in promoting real convergence of these
economies to EU standards. Furthermore, given the difficulties of data gathering one would encourage data
collection and coordination of policies and implementation agencies from early stages is critical to provide
info to help design basic programs.
Despite all the considerations above it is appropriate to conclude on a cautionary note. Although we
have established empirically the importance of public investment in transportation infrastructure for economic
development in Portugal, we have done so with a relatively small data set. This places some limitations on the
24
statistical analysis in the paper. More importantly, maybe, is the fact that establishing that public investment
has been important in the past does not establish that it will be important in the future. Indeed, one could
easily conjecture a pattern of decreasing marginal returns to public investment. Finally, even if we could
legitimately conjecture, based on the relatively high rates of return we estimated, that these public investments
will continue to be important, we did not address the issue of which types of investment are the most
important. Indeed, just showing that public investment in infrastructures is productive does not mean that it is
more productive than private investment or investment in human capital, for example.
25
REFERENCES
Aschauer, D. (1989a): “Is Public Expenditure Productive?, ” Journal of Monetary Economics, 23,
pp. 177-200.
_________ (1989b): “Does Public Capital Crowd Out Private Capital?,” Journal of Monetary
Economics, 24, pp. 171-188.
_________ (1989c): “Public Investment and Productivity Growth in the Group of Seven,” Economic
Perspectives, 13(5), pp. 17-25.
Banco de Portugal (1997) Séries Longas para a Economia Portuguesa, Lisboa, Portugal.
Bernd, E. and B. Hansson (1992). "Measuring the Contribution of Public Infrastructure in Sweden,"
Scandinavian Journal of Economics 94, pp. 151-168.
Christiano, L., M. Eichenbaum, C. Evans (1996): “The Effects of Monetary Policy Shocks: Evidence
from the Flow of Funds,” Review of Economics and Statistics 78(1), pp. 16-34.
_________ (1998): “Monetary Policy Shocks: What Have we Learned and to What End?,” NBER
Working Paper 6400.
Commission of the European Communities (1999): European Economy No 69 - Statistical Annex,
Brussels, Belgium.
Evans and Karras (1993): “Is Government Capital Productive? Evidence From a Panel of Seven
Countries,” Journal of Macroeconomics, 16(2), pp. 271-279.
Ford, R. and P. Poret (1991). "Infrastructure and Private Sector Productivity," OECD Economic
Studies 17, pp. 63-89.
Gonzalo, J, and T. Lee (1998). "Pitfalls in Testing for Long-Run Relationships," Journal of
Econometrics 86, pp. 129 - 154.
Gramlich, E. (1994): “Infrastructure Investment: A Review Essay,” Journal of Economic Literature,
XXXII(3), pp. 1177 - 1196.
Hulten C. and R. Schwab (1993): “Infrastructure Spending: Where Do We Go From Here?,”
National Tax Journal, XLVI, pp. 261 - 274.
Jorgenson, D. (1991): “Fragile Statistical Foundations,” The Public's Capital, pp. 6-7.
Lynde and Richmond (1993): “Public Capital and Long-Run Costs in U.K. Manufacturing,” The
Economic Journal, 103, pp. 880-893.
Merriman, D. (1990): “Public Capital and Regional Output: Another Look at Some Japanese and
American Data,” Regional Science and Urban Economics, 20, pp. 437-458.
Ministério das Finanças (2000), The Portuguese Economy, Direcção Geral de Estudos e Previsão,
Lisbon, Portugal.
26
Mittnik, S, and T. Newman (1998): “Dynamic Effects of Public Investment,” Institute of Statistics
and Econometrics Christian Albrechts, University at Kiel, Olshauserstr, Germany.
Munnell, A. (1992): “Infrastructure Investment and Economic Growth,” Journal of Economic
Perspectives, 6, pp. 189 -198.
Otto, G. and G. Voss (1996): “Public Capital and Private Production in Australia,” Southern
Economic Journal, 62(3), pp. 723-738.
Pereira, A. (2000): “Is All Public Capital Created Equal?,” Review of Economics and Statistics, 82
(3), pp. 513-518.
Pereira A. and J. Andraz (2001): “Investimento Público em Infra-estruturas de Transporte em
Portugal Continental,” Ministério do Planeamento, Portugal.
Pereira, A. and O. Roca (1999): “Public Capital Formation and Regional Development in Spain,”
Review of Development Economics, 3(3), pp. 281-294.
Reich, R. (1991): “The Real Economy,” The Atlantic Monthly, pp. 35-52.
Rudebusch, G. D. (1998): “Do Measures of Monetary Policy in a VAR Make Sense?,” International
Economic Review, 39(4), pp.907-931.
Seitz, H. (1994): “Public Capital and the Demand for Private Inputs,” Journal of Public Economics,
54, pp. 287-307.
Shah, A. (1992): “Dynamics of Public Infrastructure, Industrial Productivity and Profitability,” The
Review of Economics and Statistics, LXXIV(1), pp. 28-36.
Sturm, J. and J. de Haan (1995). Is Public Expenditure Really Productive? New
Evidence for the USA and the Netherlands," Economic Modeling 12, pp. 60-72.
27
Table 1 - Data set for Portugal
Years
Output
Employment
Private investment
Public investment
1976
9.023.294
3.624
2.033.225
113.212
1977
9.519.412
3.671
2.265.797
103.305
1978
9.788.215
3.770
2.427.918
123.781
1979
10.341.059
3.862
2.376.059
134.398
1980
10.813.624
3.943
2.577.030
122.724
1981
10.991.600
3.939
2.728.533
159.585
1982
11.224.467
3.965
2.770.359
162.976
1983
11.205.164
3.878
2.579.592
137.623
1984
10.991.821
3.937
2.137.447
114.935
1985
11.305.010
3.932
2.071.879
127.688
1986
11.768.296
3.900
2.293.706
144.316
1987
12.518.244
4.006
2.677.251
155.751
1988
13.451.070
4.096
3.102.701
178.492
1989
14.144.498
4.236
3.236.613
198.425
1990
14.759.562
4.279
3.501.980
247.907
1991
15.104.553
4.335
3.624.977
295.401
1992
15.483.749
4.359
3.801.235
269.922
1993
15.311.404
4.295
3.585.621
306.278
1994
15.651.492
4.449
3.701.556
324.773
1995
16.102.000
4.416
3.880.582
311.933
1996
16.584.869
4.445
4.067.238
335.804
1997
17.260.447
4.530
4.545.614
433.520
1998
17.866.194
4.740
4.952.094
404.936
Units:
Output, private investment, and public investment – millions of 1995 escudos.
Labor – thousand workers.
Sources:
Output, employment and private investment: the Bank of Portugal/Banco de Portugal (1997),
Commission of the European Communities (1999), and Ministry of Finance/Ministério das Finanças
(2000)
Public investment: Pereira and Andraz (2001).
28
Table 2 - Public investment in transportation infrastructures
Aggregate
public
investment
National
Municipal
Highways
Ports
Airports
Railways
roads
roads
1976
113.212
28.538
20.900
16.216
7.110
3.365
37.083
1977
103.305
20.773
24.829
12.310
8.476
2.595
34.322
1978
123.781
26.976
33.694
16.570
13.350
1.535
31.657
1979
134.398
30.532
49.580
9.823
14.878
2.411
27.174
1980
122.724
38.032
29.789
16.553
24.567
1.359
12.424
1981
159.585
41.197
49.274
25.381
25.381
3.442
14.911
1982
162.976
40.038
50.851
19.051
24.853
6.567
21.616
1983
137.623
43.451
41.734
9.760
21.032
5.078
16.568
1984
114.935
32.908
43.160
1.283
17.412
6.839
13.333
1985
127.688
35.674
43.418
11.159
15.543
4.469
17.425
1986
144.316
45.205
39.258
10.393
16.489
10.473
22.497
1987
155.751
52.960
43.753
13.288
12.834
8.702
24.215
1988
178.492
55.360
51.542
18.243
12.324
11.686
29.336
1989
198.425
56.892
57.549
35.603
10.185
11.041
27.156
1990
247.907
78.109
56.034
56.980
13.917
10.923
31.945
1991
295.401
84.248
67.958
67.331
18.409
8.936
48.518
1992
269.922
100.680
74.192
24.851
13.544
5.718
50.937
1993
306.278
100.777
83.302
43.571
13.326
5.207
60.095
1994
324.773
119.079
66.098
59.462
12.297
5.336
62.500
1995
311.933
128.160
54.440
57.847
12.730
6.146
52.610
1996
335.804
139.065
71.566
37.089
9.983
8.126
69.975
1997
433.520
126.036
90.559
75.553
12.059
10.502
118.811
1998
404.936
118.974
93.988
52.518
16.675
14.217
108.563
Years
Units:
Millions of 1995 escudos.
Source: Pereira and Andraz (2001).
29
Table 3 – Public investment as a share of GDP (%)
Public investment
1976-88
Averages
1989-98
Sample
Average
1976-80
1981-85
1986-88
1989-93
1994-98
pinv: Aggregate public investment
1.21
1.26
1.27
1.76
2.16
1.24
1.96
1.55
pinv1: National roads
0.29
0.35
0.41
0.56
0.76
0.34
0.66
0.48
pinv2: Municipal roads
0.32
0.41
0.36
0.45
0.45
0.36
0.45
0.40
pinv3: Highways
0.15
0.12
0.11
0.31
0.34
0.13
0.32
0.21
pinv4: Ports
0.14
0.19
0.11
0.09
0.08
0.15
0.08
0.12
pinv5: Airports
0.02
0.05
0.08
0.06
0.05
0.05
0.05
0.05
pinv6: Railways
0.30
0.15
0.20
0.29
0.49
0.22
0.39
0.29
1976-88
Averages
1989-98
Table 4 – Public investment as a share of Private Investment (%)
Public investment
Sample
Average
1976-80
1981-85
1986-88
1989-93
1994-98
pinv: Aggregate public investment
5.13
5.72
5.95
7.40
8.56
5.55
7.98
6.60
pinv1: National roads
1.24
1.58
1.91
2.35
3.02
1.53
2.69
2.03
pinv2: Municipal roads
1.35
1.87
1.67
1.91
1.77
1.63
1.84
1.72
pinv3: Highways
0.62
0.52
0.51
1.29
1.35
0.55
1.32
0.89
pinv4: Ports
0.57
0.84
0.53
0.39
0.30
0.67
0.35
0.53
pinv5: Airports
0.10
0.22
0.39
0.24
0.20
0.21
0.22
0.22
pinv6: Railways
1.25
0.69
0.94
1.22
1.91
0.96
1.57
1.23
1976-80
1981-85
1986-88
1989-93
1994-98
1976-88
Averages
1989-98
pinv1: National roads
24.2
27.7
32.1
31.8
35.5
27.4
33.7
30.1
pinv2: Municipal roads
26.2
32.8
28.1
25.9
20.6
29.2
23.3
26.6
pinv3: Highways
12.1
8.9
8.7
17.4
15.7
10.1
16.5
12.9
pinv4: Ports
11.3
14.8
8.9
5.3
3.5
12.1
4.4
8.7
pinv5: Airports
1.9
3.9
6.5
3.4
2.4
3.7
2.9
3.4
pinv6: Railways
24.4
12.0
15.9
16.3
22.2
17.6
19.3
18.3
Table 5 - Shares of total public investment (%)
Public investment
Sample
Average
30
Share of Private Investment
Share of GDP
98
97
19
96
19
95
Years
19
94
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
19
93
(% of Private Investment)
0.00
92
0.00
91
2.00
90
0.50
89
4.00
88
1.00
87
6.00
86
1.50
84
85
8.00
82
83
2.00
81
10.00
79
80
2.50
77
78
12.00
76
3.00
19
(% of GDP)
Figure 1: Aggregate public investment in transportation infrastructures (pinv)
as % of GDP and private investment
31
0.90
4.00
0.80
3.50
0.70
3.00
(% of GDP)
0.60
2.50
0.50
2.00
0.40
1.50
0.30
1.00
0.20
0.00
0.00
19
79
19
80
19
81
19
82
19
83
19
84
19
85
19
86
19
87
19
88
19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97
19
98
19
19
19
77
78
0.50
76
0.10
(% of Private Investment)
Figure 2: Public investment in national roads (pinv1)
as % of GDP and private investment
Years
Share of GDP
Share of Private Investment
Figure 3: Public investment in municipal roads (pinv2)
as % of GDP and private investment
2.50
0.60
0.50
1.50
0.30
1.00
0.20
0.50
0.10
19
79
19
80
19
81
19
82
19
83
19
84
19
85
19
86
19
87
19
88
19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97
19
98
78
19
19
77
0.00
76
0.00
19
(% of GDP)
0.40
Years
Share of GDP
Share of Private Investment
(% of Private Investment)
2.00
32
0.45
1.80
0.40
1.60
0.35
1.40
0.30
1.20
0.25
1.00
0.20
0.80
0.15
0.60
0.10
0.40
0.05
0.20
0.00
0.00
19
79
19
80
19
81
19
82
19
83
19
84
19
85
19
86
19
87
19
88
19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97
19
98
19
19
19
(% of Private Investment)
2.00
77
78
0.50
76
(% of GDP)
Figure 4: Public investment in highways (pinv3)
as % of GDP and private investment
Years
Share of GDP
Share of Private Investment
Figure 5: Public investment in ports (pinv4)
as % of GDP and private investment
0.25
1.20
0.60
0.10
0.40
0.05
0.20
19
79
19
80
19
81
19
82
19
83
19
84
19
85
19
86
19
87
19
88
19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97
19
98
78
19
19
77
0.00
76
0.00
19
(% of GDP)
0.80
0.15
Years
Share of GDP
Share of Private Investment
(% of Private Investment)
1.00
0.20
33
0.09
0.45
0.08
0.40
0.07
0.35
0.06
0.30
0.05
0.25
0.04
0.20
0.03
0.15
0.02
0.10
0.01
0.05
0.00
0.00
19
79
19
80
19
81
19
82
19
83
19
84
19
85
19
86
19
87
19
88
19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97
19
98
19
19
19
(% of Private Investment)
0.50
77
78
0.10
76
(% of GDP)
Figure 6: Public investment in airports (pinv5)
as % of GDP and private investment
Years
Share of GDP
Share of Private Investment
Figure 7: Public investment in railways (pinv6)
as % of GDP and private investment
3.00
0.80
0.70
2.00
0.50
0.40
1.50
0.30
1.00
0.20
0.50
0.10
0.00
76
19
77
19
78
19
79
19
80
19
81
19
82
19
83
19
84
19
85
19
86
19
87
19
88
19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97
19
98
0.00
19
(% of GDP)
0.60
Years
Share of GDP
Share of Private Investment
(% of Private Investment)
2.50
34
Figure 8: Composition of public investment in transportation infrastructures
1.00
0.80
0.60
0.40
0.20
0.00
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
Years
National roads
Municipal roads
Highways
Ports
Airports
Railways
35
Figure 9: Aggregate public investment in transportation infrastructures
as % of GDP in Portugal, Spain and in the U.S.A.
3.00
2.50
1.50
1.00
0.50
0.00
19
76
19
77
19
78
19
79
19
80
19
81
19
82
19
83
19
84
19
85
19
86
19
87
19
88
19
89
19
90
19
91
19
92
19
93
19
94
19
95
19
96
19
97
19
98
(%)
2.00
Years
Spain
U.S.A.
Portugal
36
Table 6 – Testing for the null hypothesis of unit roots using the ADF test
Deterministic
components
Order
(BIC)
Test statistic
Critical values
5%
1%
gdp: Output
CT
2
-3.0118
-3.60
-4.38
emp: Employment
CT
0
-1.6349
-3.60
-4.38
inv: Private investment
CT
1
-2.0732
-3.60
-4.38
pinv: Aggregate public investment
CT
0
-1.8699
-3.60
-4.38
pinv1: National roads
CT
0
-1.9461
-3.60
-4.38
pinv2 : Municipal roads
CT
1
-3.2594
-3.60
-4.38
pinv3: Highways
CT
0
-3.4357
-3.60
-4.38
C
0
-2.1310
-3.00
-3.75
pinv5: Airports
CT
2
-2.7925
-3.60
-4.38
pinv6. Railways
CT
2
-0.9072
-3.60
-4.38
gdp: Output
C
3
-4.6470
-3.00
-3.75
emp: Employment
C
0
-3.9231
-3.00
-3.75
inv: Private investment
N
0
-2.2950
-1.95
-2.66
pinv: Aggregate public investment
C
0
-4.8404
-3.00
-3.75
pinv1: National roads
C
0
-5.5086
-3.00
-3.75
pinv2 : Municipal roads
C
1
-4.1368
-3.00
-3.75
pinv3: Highways
N
0
-4.9659
-1.95
-2.66
pinv4: Ports
N
0
-3.3923
-1.95
-2.66
pinv5: Airports
N
0
-4.7955
-1.95
-2.66
pinv6. Railways
CT
1
-5.8943
-3.60
-4.38
Variables: log levels
pinv4: Ports
Variables: growth rates
37
Table 7 – Testing for the null hypothesis of unit roots using the Phillips-Perron test
Deterministic
components
Order
(BIC)
Test statistic
Critical values
5%
1%
C
1
-0.1828
-12.5
-17.2
emp: Employment
CT
0
-8.3216
-17.9
-22.5
inv: Private investment
CT
1
-22.4939
-17.9
-22.5
pinv: Aggregate public investment
CT
0
-8.0042
-17.9
-22.5
pinv1: National roads
CT
0
-11.7611
-17.9
-22.5
pinv2 : Municipal roads
CT
1
-17.4691
-17.9
-22.5
pinv3: Highways
CT
0
-13.6992
-17.9
-22.5
C
0
-8.3041
-12.5
-17.2
pinv5: Airports
CT
2
-9.1929
-17.9
-22.5
pinv6. Railways
CT
2
-6.0600
-17.9
-22.5
gdp: Output
C
0
-10.1261
-12.5
-17.2
emp: Employment
C
0
-21.0370
-12.5
-17.2
inv: Private investment
N
0
-8.4164
-7.3
-11.9
pinv: Aggregate public investment
C
0
-23.0084
-12.5
-17.2
pinv1: National roads
C
0
-23.3204
-12.5
-17.2
pinv2 : Municipal roads
C
1
-36.6651
-12.5
-17.2
pinv3: Highways
N
0
-23.1967
-7.3
-11.9
pinv4: Ports
N
0
-16.1431
-7.3
-11.9
pinv5: Airports
N
0
-22.5798
-7.3
-11.9
pinv6. Railways
CT
1
-113.4204
-17.9
-22.5
Variables: log levels
gdp: Output
pinv4: Ports
Variables: growth rates
38
Table 8 – Testing the null hypothesis of no cointegration
Variables
Deterministic
Components
Optimal Lag
(BIC)
Test Statistic
CT
0
-1.4347
-4.16
-4.65
C
0
-3.9402
-4.11
-4.73
inv: Private investment
CT
0
-3.1888
-4.16
-4.65
pinv: Aggregate public investment
CT
0
-4.5880
-4.16
-4.65
gdp: Output
CT
0
-1.3356
-4.16
-4.65
C
0
-3.9385
-4.11
-4.73
inv: Private investment
CT
1
-2.8594
-4.16
-4.65
pinv1: National roads
N
2
-3.5147
-3.74
-4.30
gdp: Output
CT
0
-1.4802
-4.16
-4.65
C
0
-4.3649
-4.11
-4.73
inv: Private investment
CT
0
-2.0492
-4.16
-4.65
pinv2 : Municipal roads
C
1
-4.8360
-4.11
-4.73
CT
0
-2.1859
-4.16
-4.65
C
0
-4.0019
-4.11
-4.73
inv: Private investment
CT
0
-1.9255
-4.16
-4.65
pinv3: Highways
CT
1
-5.3735
-4.16
-4.65
gdp: Output
CT
0
-1.7135
-4.16
-4.65
C
2
-5.2319
-4.11
-4.73
inv: Private investment
CT
0
-1.6549
-4.16
-4.65
pinv4: Ports
C
0
-3.3592
-4.11
-4.73
gdp: Output
CT
0
-1.3165
-4.16
-4.65
C
0
-4.0143
-4.11
-4.73
inv: Private investment
CT
0
-1.8647
-4.16
-4.65
pinv5: Airports
N
2
-2.9691
-3.74
-4.70
gdp: Output
CT
0
-2.0721
-4.16
-4.65
C
0
-4.0632
-4.11
-4.73
inv: Private investment
CT
0
-3.6027
-4.16
-4.65
pinv6. Railways
CT
0
-3.7113
-4.16
-4.65
gdp: Output
emp: Employment
emp: Employment
emp: Employment
gdp: Output
emp: Employment
emp: Employment
emp: Employment
emp: Employment
Critical Values
5%
1%
39
Table 9 – VAR specification (BIC)
Public investment
pinv: Aggregate public investment
pinv1: National roads
pinv2 : Municipal roads
pinv3: Highways
pinv4: Ports
pinv5: Airports
pinv6. Railways
Model
order
Deterministic
components
No dummy
One dummy
(1989)
Two dummies
(1989,1994)
1
N
-24.79172
-24.89442
-25.11975
1
C
-25.05496
-25.46241
-25.69166
1
CT
-25.09007
-25.58873
-25.69035
1
N
-24.85156
-24.94396
-25.52282
1
C
-25.07617
-25.52986
-26.10160
1
CT
-25.24993
-26.45380
-26.77999
1
N
-23.64226
-23.68089
-23.73939
1
C
-23.92128
-24.30213
-24.36215
1
CT
-24.11397
-24.51930
-24.59061
1
N
-20.04118
-20.08434
-20.13513
1
C
-20.32396
-20.75705
-20.81506
1
CT
-20.37029
-20.87376
-21.01449
1
N
-23.77836
-23.84164
-23.89150
1
C
-24.06601
-24.45286
-24.51007
1
CT
-24.31672
-24.93027
-25.28925
1
N
-21.64709
-21.68944
-21.81481
1
C
-21.96905
-22.52255
-22.66662
1
CT
-22.11673
-22.50320
-22.62232
1
N
-22.95706
-23.30818
-23.35772
1
C
-23.47973
-23.91985
-23.97603
1
CT
-23.57522
-23.87059
-23.92901
NB: In bold face is the selected specification.
40
Table 10 – Variance decomposition: percentage of long-term variation in the variables due to variations in public investment
Variable
Output
Employment
Investment
Public Investment
pinv: Aggregate public investment
central case
range of variation
37.6%
[9.8%;37.6%]
18.7%
[8.8%;19.2%]
37.2%
[6.9%;37.2%]
88.7%
[69.2%;88.7%]
central case
range of variation
33.5%
[0.2%;35.6%]
3.3%
[0.6%;24.8%]
43.9%
[0.1%;45.3%]
76.7%
[24.9%;76.7%]
central case
range of variation
11.2%
[3.8%;11.2%]
11.0%
[7.9%;11.0%]
5.8%
[1.8%;5.8%]
64.1%
[57.9%;64.1%]
central case
range of variation
7.8%
[0.9%;7.8%]
3.9%
[2.5%;3.9%]
17.0%
[2.2%;17.0%]
53.7%
[38.4%;55.0%]
central case
range of variation
16.7%
[0.8%;16.7%]
32.1%
[1.9%;32.1%]
12.8%
[0.5%;12.8%]
65.7%
[28.7%;65.7%]
central case
range of variation
0.6%
[0.6%;5.5%]
7.1%
[0.4%;9.4%]
6.1%
[0.9%;11.6%]
90.5%
[80.7%;90.5%]
central case
range of variation
12.4%
[0.9%;23.4%]
10.8%
[0.7%;16.1%]
22.3%
[2.9%;30.5%]
90.5%
[64.4%;90.5%]
pinv1: National roads
pinv2: Municipal roads
pinv3: Highways
pinv4: Ports
pinv5: Airports
pinv6: Railways
41
Table 11 – Long-term accumulated elasticities of private sector variables with respect to public investment
Variable
Output
Employment
Investment
central case
range of variation
0.18264
[0.105;0.183]
0.07860
[0.045;0.079]
0.63871
[0.356;0.639]
central case
range of variation
0.19807
[-0.133;0.202]
0.04524
[-0.206;0.049]
0.76549
[-0.259;0.772]
central case
range of variation
0.09839
[0.054;0.098]
0.05441
[0.032;0.054]
0.25396
[0.111;0.254]
central case
range of variation
0.02416
[0.000;0.024]
0.00865
[0.000;0.009]
0.11013
[0.010;0.110]
central case
range of variation
0.08736
[-0.057;0.087]
0.07025
[0.005;0.070]
0.28102
[-0.075;0.281]
central case
range of variation
0.00937
[-0.014;0.030]
-0.00438
[-0.005;0.009]
0.07858
[0.002;0.137]
central case
range of variation
0.06247
[0.014;0.090]
0.01221
[-0.010;0.031]
0.26418
[0.080;0.341]
pinv: Aggregate public investment
pinv1: National roads
pinv2: Municipal roads
pinv3: Highways
pinv4: Ports
pinv5: Airports
pinv6: Railways
0.01652
(0.22)
0.08439
(0.79)
0.43479
(2.37)**
-0.85095
(-0.74)
0.19952
(1.00)
0.60145
(1.97)**
0.12489
(0.75)
pinv: Aggregate public investment
pinv1: National roads
pinv2: Municipal roads
pinv3: Highways
pinv4: Ports
pinv5: Airports
pinv6: Railways
NB: t-statistics in parenthesis.
* Significant at 10% level.
** Significant at 5% level.
Constant
GPINV
----
----
-0.05471
(-2.69)**
0.24811
(2.13)**
-0.02938
(-1.62)*
-0.01174
(-1.10)
----
Trend
Table 12 – Policy functions for different types of public investment
0.17389
(1.00)
-0.60197
(-1.91)**
0.54009
(2.68)**
-2.78687
(-2.25)**
0.16555
(0.86)
0.20328
(1.76)**
0.08141
(1.02)
D1989
0.17051
(0.87)
-0.10877
(-0.32)
0.82740
(2.98)**
-3.82736
(-2.21)**
0.28084
(1.04)
0.08463
(0.53)
0.00836
(0.10)
D1994
-3.33996
(-0.54)
-8.35899
(-0.77)
9.67703
(2.40)**
-0.53118
(-0.02)
-5.38487
(-1.28)
3.98963
(1.59)*
1.06899
(0.38)
GGDP(-1)
-3.75863
(-0.69)
-4.37766
(-0.45)
0.24936
(0.07)
53.91315
(2.19)**
-0.20668
(-0.05)
0.98707
(0.43)
0.91162
(0.37)
GEMP(-1)
1.32406
(0.79)
2.86747
(1.00)
-1.66084
(-1.46)*
-15.56848
(-1.61)*
2.93549
(2.63)**
-0.78891
(-1.16)
0.43275
(0.55)
GINV(-1)
-0.12572
(-0.46)
-0.38275
(-1.57)*
-0.12838
(0.56)
-0.29846
(-2.25)**
-0.34957
(-1.72)
-0.18631
(-0.96)
-0.31825
(-1.18)
GPINV(-1)
43
44
Figure 10: Accumulated impulse response function for aggregate public investment (pinv)
1.4
1.2
1
0.8
0.6
0.4
0.2
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Years
Output
Employment
Private investment
Aggregate public investment (pinv)
45
Figure 11: Accumulated impulse response functions for national roads (pinv1)
1.2
1
0.8
0.6
0.4
0.2
0
0
1
2
3
4
5
6
Output
7
8
9
Employment
10 11
Years
12
13
Private investment
14
15
16
17
18
19
20
18
19
20
National roads
Figure 12: Accumulated impulse response functions for municipal roads (pinv2)
1.2
1
0.8
0.6
0.4
0.2
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Years
Output
Employment
Private investment
Municipal roads (pinv2)
46
Figure 13: Accumulated impulse response functions for Highways (pinv3)
1.2
1
0.8
0.6
0.4
0.2
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Years
Output
Employment
Private investment
Highways (pinv3)
Figure 14: Accumulated impulse response functions for ports (pinv4)
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Years
Output
Employment
Private investment
Ports (pinv4)
47
Figure 15: Accumulated impulse response functions for airports (pinv5)
1.2
1
0.8
0.6
0.4
0.2
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
-0.2
Years
Output
Employment
Private investment
Airports (pinv5)
Figure 16: Accumulated impulse response functions for railways (pinv6)
1.2
1
0.8
0.6
0.4
0.2
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Years
Output
Employment
Private investment
Railways (pinv6)
48
Table 13 – Effects of public investment on private investment
Elasticities
Public investment variable
Marginal productivity
pinv: Aggregate public investment
central case
0.63871
8.12
central case
0.76549
29.58
central case
0.25396
14.05
central case
0.11013
9.19
central case
0.28102
84.40
central case
0.07858
39.13
central case
0.26418
18.83
pinv1: National roads
pinv2: Municipal roads
pinv3: Highways
pinv4: Ports
pinv5: Airports
pinv6: Railways
49
Table 14 – Effects of public investment on employment
Elasticities
Public investment variable
Number of jobs
(per million of Euros)
pinv: Aggregate public investment
central case
0.07860
230
central case
0.04524
404
central case
0.05441
692
central case
0.00865
164
central case
0.07025
4800
central case
-0.00438
-500
central case
0.01221
204
pinv1: National roads
pinv2: Municipal roads
pinv3: Highways
pinv4: Ports
pinv5: Airports
pinv6: Railways
50
Table 15 - Effects of public investment on output
Elasticities
Marginal productivity
central case
0.18264
9.54
15.9%
central case
0.19807
31.41
23.0%
central case
0.09839
22.32
20.9%
central case
0.02416
8.24
15.0%
central case
0.08736
107.14
30.8%
central case
0.00937
19.18
20.0%
central case
0.06247
18.47
19.7%
Public investment variable
Rates of return
pinv: Aggregate public investment
pinv1: National roads
pinv2: Municipal roads
pinv3: Highways
pinv4: Ports
pinv5: Airports
pinv6: Railways